Learning to Translate Sequence and Structure to Function: Identifying DNA Binding and Membrane Binding Proteins

Abstract

Abstract-A proteinÕs function depends in a large part on interactions with other molecules. With an increasing number of protein structures becoming available every year, a corresponding structural annotation approach identifying such interactions grows more expedient. At the same time, machine learning has gained popularity in bioinformatics providing robust annotation of genes and proteins without sequence homology. Here we have developed a general machine learning protocol to identify proteins that bind DNA and membrane. In general, there is no theory or even rule of thumb to pick the best machine learning algorithm. Thus, a systematic comparison of several classification algorithms known to perform well is investigated. Indeed, the boosted tree classifier is found to give the best performance, achieving 93% and 88% accuracy to discriminate non-homologous proteins that bind membrane and DNA, respectively, significantly outperforming all previously published works. We also attempted to address the importance of the attributes in function prediction and the relationships between relevant attributes. A graphical model based on boosted trees is applied to study the important features in discriminating DNA-binding proteins. In summary, the current protocol identified physical features important in DNA and membrane binding, rather than annotating function through sequence similarity.

Heterodimer Binding Scaffolds Recognition via the Analysis of Kinetically Hot Residues

“…To fully comprehend the biological process on the molecular level we first have to understand the physical laws that govern the interactions of biological polymers. Protein-DNA and protein-lipid interactions had been successfully addressed [ 9 , 10 , 11 , 12 , 13 ], but the problems of protein folding [ 14 , 15 ] and protein-protein interactions [ 16 , 17 , 18 , 19 , 20 , 21 ] are issues that still require the full attention of the research community. Many attempts were made to develop a comprehensive protein-protein interaction theory.…”

The structure and interaction mechanism of a polyelectrolyte complex: a dissipative particle dynamics study

“…The development of polyelectrolyte complexes as biomaterials has theoretical and experimental interest because the complexation of proteins with polyelectrolytes is the basis of processes such as protein purification, enzyme immobilization, immunosensing, and the design of bioactive sensors 13,14 . Studies of polyelectrolyte complexes have also allowed to understand the behavior of some biological macromolecules, such as DNA-binding proteins 15,16 ; in particular, Kabanov et al have used DNA-polycation complexes for the delivery of genetic material into cells, i.e., for gene transfer and gene therapy 17 . The use of polymers in gene therapy systems is mainly motivated by their specific properties such as biodegradability 18 , biocompatibility 19 , and bioactivity 20 .…”

Functional Site Discovery From Incomplete Training Data: A Case Study With Nucleic Acid–Binding Proteins

“…The characteristics of each dataset are summarized in Table 1 . Both datasets have been used in previous studies to identify residues that bind DNA (Szilagyi and Skolnick, 2006; Langlois et al, 2007) and RNA (Terribilini et al, 2006; Langlois et al, 2007; Kumar et al, 2008). During training, each residue in a DNA-binding protein is considered DNA-binding and in a non-DNA-binding protein non-binding during training and cross-validation.…”

“…These proteins play an essential role in nearly every cellular process. A number of experimental (Cajone et al, 1989; Freeman et al, 1995; Chou et al, 2003; Buck and Lieb, 2004; Nutiu et al, 2011; Gordan et al, 2013) and computational (Bhardwaj et al, 2005; Szilagyi and Skolnick, 2006; Bhardwaj and Lu, 2007; Langlois et al, 2007; Tjong and Zhou, 2007; Gao and Skolnick, 2009; Langlois and Lu, 2010b; Weirauch et al, 2013; Xu et al, 2015) approaches have been developed to identify these proteins and their functional sites. Since DNA- and RNA-binding proteins provide a substantial number of labeled examples, e.g., residues known to bind DNA or RNA, these problems have been studied extensively thus presenting an excellent proof of concept for our approach.…”